An enhanced harmonic densely connected hybrid transformer network architecture for chronic wound segmentation utilising multi-colour space tensor merging

<p>Chronic wounds and associated complications pose an ever-increasing burden on clinics and hospitals worldwide. Venous, arterial, diabetic, and pressure ulcers are becoming increasingly common globally. These conditions can have severely debilitating consequences for those affected, with limb amputations and an increased risk of mortality due to infection becoming more common. New methods to assist clinicians in chronic wound care are therefore vital to maintaining high-quality care standards. This paper presents an improved HarDNet segmentation architecture that integrates a contrast-eliminating component into the initial layers of the network to enhance feature learning. We also utilize a multi-color space tensor merging process and adjust the harmonic shape of the convolution blocks to facilitate the extraction of these additional features. We train our proposed model using wound images from light-skinned patients and test the model on two test sets (one with ground truth and one without) comprising only darker-skinned cases. Subjective ratings are obtained from clinical wound experts, with the intraclass correlation coefficient used to determine inter-rater reliability. For the dark-skinned test set with ground truth, when comparing the baseline results (DSC=0.6389, IoU = 0.5350) with the results for the proposed model (DSC = 0.7610, IoU = 0.6620), we demonstrate improvements in terms of the Dice similarity coefficient (+0.1221) and intersection over union (+0.1270). Measures from the qualitative analysis also indicate improvements in terms of high expert ratings, with improvements of >3% demonstrated when comparing the baseline model with the proposed model. This paper presents the first study to focus on darker skin tones for chronic wound segmentation using models trained only on wound images exhibiting lighter skin. Diabetes is highly prevalent in countries where patients have darker skin tones, highlighting the need for a greater focus on such cases. Additionally, we conduct the largest qualitative study to date on chronic wound segmentation. All source code for this study is available at: https://github.com/mmu-dermatology-research/hardnet-cws.</p>

DOI 10.1016/j.compbiomed.2025.110172

Arterial ulcers

Chronic wounds

Deep learning

Diabetic foot ulcers

Hybrid transformer

Pressure ulcers

Segmentation

Synthetic wounds

Venous ulcers

Wounds analysis